physical constants

I'm in the middle of editing the physical constants
in Annex D.  Per this morning's conference call,
since most vendors have left P_K, P_Q, etc. at the
values defined in the 2.0 LRM, we are going to
revert those values.

The values in the 2.1 and 2.2 LRM were taken from
physics.nist.gov during the 2.1 LRM update.

I have come across one quandary: in LRM 2.0, we have
`define P_C 2.997924562e8
However, NIST now says that the value 
`define P_C 2.99792458e8
is exact.  (I believe the meter is now defined in
terms of how far light travels in 1/c of a second.)

Also, LRM 2.0 says
`define P_EPS0 8.85418792394420013968e-12
whereas NIST now says
`define P_EPS0 8.854187817e-12
is exact.

(Similarly, P_U0 and P_CELSIUS0 are exact.)

I propose that, for those values that NIST now says
are exact, there should be only one value in 
constants.vams, despite the fact that older LRMs
had different values.  That leaves P_Q, P_K, and
P_H with differing values.

-Geoffrey